9. Epidemiology of surgical patients

Much has been written about Dupuytren's disease and many opinions have been expressed, often on the basis of only a few cases. Many of the statistics in use and commonly cited are based upon hospital patients and in many respects they show discrepancies because these patients are highly selected, and some bias can hardly be avoided. In groups of patients with primary diseases such as alcoholism or epilepsy, the related incidence of Dupuytren's disease is high but these patients do not necessarily seek surgical intervention for their contracture. On the other hand, patients with Dupuytren's disease who consult the surgeon do not reflect the features of the disease in the general population where many people have minimal disease, elderly patients often have it without their knowledge, and others are content to accept contracture or have been advised not to have an operation.

The epidemiological pattern of patients examined and in most instances operated upon for Dupuytren's disease have been studied with three purposes:

to obtain objective data on the type of Dupuytren's disease a surgeon can be faced with in Belgium;
to compare those data with the literature to make sure that the manifestation of the disease is the same as that seen in other countries so as to be able to compare the surgical results;
to analyse the relationships between the factors supposed to play a role in the severity of the disease and verify if there is some statistical evidence in support of the concept of a Dupuytren diathesis.

9.1 The patients

The study population consists of 538 patients with Dupuytren's disease who consulted specifically about this problem between 1983 and 1994. Patients who came for another condition like a carpal tunnel syndrome or trigger fingers and in whom Dupuytren's nodules where found on examination, are not included in this survey.

The study started with a prospective evaluation of segmental aponeurectomy in 1983. At the beginning, only those cases operated upon with this technique were included. Later on, all operated cases were included and also, progressively, all patients who consulted about a Dupuytren's disease.

Several factors that could play a role in the development or the progression of the disease have been studied: age, sex, age of onset, length of evolution before the first treatment or consultation, hand dominance, first hand affected, manual labour, family history, personal medical history, ectopic sites, unilateral or bilateral involvement and previous operation.

9.2 Age and sex differences

Of the 538 patients examined 414 (77 %) were men and 124 (23 %) women. The youngest man was 24 and the oldest 87. In women the corresponding ages were 29 and 88 years. The mean age for all patients was 60.2, for men 59.6 (STD 11.2) and for women 62.4 (STD 10.6) years. This difference is statistically significant with p < 0.02 (statistical methodology: two-sample pooled-variance t test; Levene's test was used to test the hypothesis that the two population variances were equal (SPSS¹).

The age distribution in both sexes is shown in figures 1 and 2. Most patients were between 50 and 70 years old. The ratio of men to women changed with the age groups and seemed to decrease gradually after the age of 40 (fig. 3).


	Figure 9-1: Age distribution. The percentages are calculated for each sex.
	Figure 9-2: Age distribution. The percentages are calculated for the whole population.
	Figure 9-3: Ratio of men to women by age group

9.2.1 Length of evolution

All patients were asked since how long the disease had been progressing. This measurement is certainly not very reliable because, as we all know, the recollection of elapsed time is at best imprecise. Nevertheless it provides some interesting information.

The mean length of evolution between the onset of the disease and the first consultation for the whole study population was 6.8 years. For men it was 7.3 (S.D. 6.4) and for women 5.4 years (S.D. 4.9). This difference is statistically significant with p < 0.005 (statistical methodology: the length of evolution does not have a normal distribution; the difference between the two groups was tested with the Mann-Whitney test which does not require assumptions about the shape of the underlying distribution (SPSS⁴)).

9.2.2 Age of onset

The age of onset was calculated by substracting the length of evolution from the age at the first consultation. As for the length of evolution this measurement is a rough estimate. The mean age of onset for men was 52.3 (S.D. 12.3) and for women 57.1 years (S.D. 10.6). This difference is statistically significant with p < 0.0005 (statistical methodology: two-sample separate-variance t test; Levene's test was used to test the hypothesis that the two population variances were equal. but they were not).

The age of onset distribution in both sexes is shown in figures 4 and 5. Most patients were between 50 and 60 years old when they noticed the first signs of the disease. The ratio of men to women changed with the age groups and clearly decreased gradually after the age of 40 (fig. 6).


	Figure 9-4: Age of onset distribution. The percentages are calculated for each sex.
	Figure 9-5: Age of onset distribution. The percentages are calculated for the whole population.
	Figure 9-6: Ratio of men to women by age of onset groups

9.2.3 Discussion

These results are in agreement with those published by other authors. They confirm that the disease is not only much more common in men (77 % of cases) but also appears significantly earlier and that men are operated upon earlier. McFarlane et al. (1990a) in their multicentric review of 1150 operated patients reports an even higher predominance of men although that percentage varies with the geographical area studied: 84 % for the whole group, 78 % in north America, 95 % in Japan, 89 % in Germany and France, 84 % in Great Britain and 76 % in Australia. Mikkelsen (1990) in his epidemiological survey of a Norwegian population found that women represent 28 % of the population showing signs of a Dupuytren's disease. Besides possible racial or genetic factors, the type of recruitment of patients probably plays a role in these observed differences.

Mikkelsen and McFarlane observed the same age differences as we did with a ratio of men to women decreasing gradually from 98 % in the age group 30-40 to 40 % in the group 80-90.

There is no explanation for the fact that women consult earlier in the course of the disease except perhaps that they have less chance of confusing a Dupuytren's disease at an early stage with the normal hardening of the palmar skin related to hard manual work.

9.3 Hand dominance and first affected side

As shown in figure 7, 91.9 % of the patients were right-handed, 3.5 % left-handed and 4.6 % ambidextrous. The left-handed patients who were taught at school to write with the right hand were coded as ambidextrous.


	Figure 9-7: Hand dominance

The disease started in the right hand in 45.6 % and in the left hand in 33.6 % of the patients. Both hands were affected simultaneously in 20.8 % of the cases (fig. 8).


	Figure 9-8

There is a relation between hand dominance and the hand which is first affected by the disease as can be seen from table 1.

This difference is nevertheless not significant (p < 0.4; statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-1: Crosstabulation of handedness and hand first affected by the disease**

9.3.1 Discussion

Baron Dupuytren himself suggested that the contracture he described was due to repetitive injuries to the hand. Since then many conflicting studies have been published on that subject. Mikkelsen (1978, 1990) found a higher prevalence of Dupuytren's disease and more severe contractures in people doing hard manual work than in those doing light or non-manual work. On the contrary, Fisk (1985) did not find that occupation had an effect upon the disease process.

If manual labour and repetitive injuries played a role in the genesis of the disease, we would expect to find a clear relation between handedness and the side of the first affected hand which is not the case in this survey. The fact that both hands were affected from the beginning in one fifth of the cases pleads against that hypothesis as well. Hueston (1990a) has made the same observation and uses this as one of the key arguments to show that the diathesis over-rules any local hand activity.

9.4 Manual labour

Occupations were noted for all patients in this study and divided into three groups: no or light, medium and heavy manual work. Groups 1 and 3 were clearly defined but the transition between groups 1 and 2 or 2 and 3 was difficult to appreciate. This introduced a part of subjectivity in the coding all the more because the activities of the patients changed frequently with time.

As shown in figure 9, 68.8 % of the patients were coded as having a light manual activity, 21.1 % a medium and 10.1 % a heavy one. In itself this distribution does not have any significance but just gives an idea of the recruitment of patients.


	Figure 9-9: Distribution of manual activity

9.4.1 Manual labour and age of onset

Is there a relation between manual activity and the age of onset of the disease? None was found as can be seen from table 2. The small observed differences in the means are not significant (p<&0.6; statistical methodology: one-way analysis of variance; Levene's test was used to test the hypothesis that the two population variances were equal (SPSS³)).

**Table 9-2: Age of onset by manual activity**

9.4.2 Discussion

If manual labour played a major role in the genesis of Dupuytren's disease, we would expect to find a relation between the age of onset of the disease and the intensity of manual activity which is not the case in this study. This is one more argument in favour of Hueston's diathesis theory.

9.5 Family history of Dupuytren's disease

It has been assumed since the study of Ling (1963) that Dupuytren's disease is genetically transmitted. By implication Hueston (1990a) has suggested that it is a disease of the Celtic race, or perhaps originated with the Vikings, because the prevalence of Dupuytren's disease coincides with the early migrations of these people as we have seen in the previous chapter. In this context, the study of the heredity of the patients is of particular interest.

All patients in this study were asked about the occurrence of the disease in their family. As shown by Ling (1963) who noted that the number of affected relatives rose from 16 % on history taking to 68 % after examination of relatives, little statistical value can be placed, from the epidemiologist point of vue, on the familial incidence as suggested by history alone. Nevertheless, this factor on which most of Hueston's theory (1990) of Dupuytren diathesis is based could have a prognostic value. Therefore, its relation with other clinical elements was studied.

A positive family history was recorded in 28.2 % of the patients (fig. 10). McFarlane et al.(1990a) in their multicentric survey, report a positive history in 29 %.


	Figure 9-10: Family history

9.5.1 Family history and sex of the patient

As seen in table 3, there is no relation between the sex of the patient and a family history of the disease (p<0.7).

**Table 9-3: Correlation of family history with the sex of the patient**

9.5.2 Family history and age of onset of the disease

Family history was correlated with age of onset (table 4).

**Table 9-4: Correlation of family history with age of onset**

The disease began on average six years earlier in patients with known relatives affected by Dupuytren's disease. This difference is highly significant (p<0.0005) and is clearly shown in figure 11 (statistical methodology: two-sample pooled-variance t test; Levene's test was used to test the hypothesis that the two population variances were equal (SPSS¹)).


	Figure 9-11: Boxplot of age of onset by family history of Dupuytren's disease. The black square approximately in the middle of the clear rectangle represents the median. The lower border of the rectangle represents the 25th percentile, the upper border the 75th percentile and so 50 % of the cases have values within the box. The lower ad upper extremities of the vertical lines represent respectively the 5th and 95th percentiles.

9.5.3 Discussion

The strong relation between family history and age of onset is in agreement with the findings of Mikkelsen (1990) who showed that the patients who developed a Dupuytren's disease at a young age had a greater incidence of positive family history. This is a strong argument for the existence of a diathesis as advocated by Hueston (1990a). McFarlane et al. (1990a) found that women knew more often than men family relatives affected by the disease. The present results are not in accordance with this.

9.6 Ectopic sites of the disease

The involvement of other areas is usually interpreted as a strong diathesis factor (McFarlane, 1990b, McFarlane et al., 1990a, Hueston 1990a).

The presence of ectopic lesions was not systematically searched for at the beginning of this study. The information was recorded in only 210 patients. The diagnosis of knuckle pads was based, as suggested by Mikkelsen (1990), upon the presence of a subcutaneous thickening over the dorsal aspect of the proximal interphalangeal joint, adherent to the skin but mobile over the joint capsule. Even with this clear clinical description, it is often difficult to be certain whether or not they are present so the recorded incidence may be incorrect. The presence of plantar fibromatosis (Leddherose's disease) and penile fibromatosis (Peyronie's disease) was established by questioning the patients.

Table 5 summarizes the findings. Knuckle pads were thus found in 18.1 % of the patients, plantar fibromatosis in 5.7 % and penile fibromatosis in 1 %. No patient showed an involvement of all three areas.

**Table 9-5: Ectopic lesions**

The proportion of patients showing signs of knuckle pads is much smaller than that recorded by Mikkelsen (1977, 1990) who found a percentage of 48.6% in men and 33.3 % in women. He insisted however that there seems to be a great geographical variation in the frequency of knuckle pads in Dupuytren's disease. McFarlane et al. (1990a) in their survey of surgical patients report a percentage much closer to our findings: 26% had signs of involvement in other areas with 20% having knuckle pads, 10% plantar fibromatosis and 2% penile fibromatosis.

9.6.1 Ectopic sites, sex of the patient and age of onset

The existence of ectopic sites of the disease was correlated with age of onset (table 6). The observed difference in the means is not significant (p<0.2).

**Table 9-6: Correlation of age of onset with the existence of ectopic lesions.**

No relation was found either with the sex of the patient (p<0.9; statistical methodology: two-sample pooled-variance t test; Levene's test was used to test the hypothesis that the two population variances were equal (SPSS¹)).

9.6.2 Ectopic sites and family history

The existence of ectopic localizations of Dupuytren's disease was correlated with family history (table 7). A difference is noticeable: only 19.1 % of the patients without family history showed signs of ectopic lesions compared with 31.3% if there was a family history. This difference is almost statistically significant (p<0.06; statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-7: Crosstabulation of family history and ectopic lesions**

9.6.3 Discussion

The absence of a clear relation between the involvement of ectopic sites and the sex of the patient, the age of onset or a positive family history is in agreement with Mikkelsen's findings (1977). It could be an argument against a Dupuytren diathesis as advocated by Hueston. The unreliability of the measure of the age of onset of the disease and of the recording of a positive family history precludes any definitive conclusion on this matter.

9.7 Involvement of both hands

Both hands were involved at the time of the first consultation in 75.2% of the cases (fig. 12). This is somewhat higher than the 65% reported by McFarlane et al. (1990a). In Mikkelsen's study (1990) both hands were involved in approximately 50% of the cases.


	Figure 9-12: Involvement of both hands

9.7.1 Involvement of both hands and ectopic lesions

The involvement of both hands was correlated with the existence of ectopic localisations of Dupuytren's disease (table 8). Ectopic lesions were present in 11.9 % of the patients with one hand affected and 30 % of the patients with both hands affected. This difference is statistically significant (p < 0.01; statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-8: Crosstabulation of ectopic lesions and number of hands affected**

9.7.2 Involvement of both hands and family history

The frequency of a positive history of Dupuytren's disease in the family is about the same in patients with one or both hands affected (table 9). A positive family history was present in 34.9% of the patients with one hand affected and oddly in only 29.6% of the patients with an involvement of both hands. This small difference is not significant (p<0.5; statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-9: Crosstabulation of history and number of hands affected**

9.7.3 Involvement of both hands and sex of the patient

No relation was found between these variables (p < 0.3).

9.7.4 Involvement of both hands and age of onset

The involvement of both hands was correlated with the age of onset. As can be seen from table 10, there is a small difference in the mean age of onset: 52.2 years if only one hand is affected and 54.2 years if both are involved. Yet this difference is not statistically significant (p<0.3).

**Table 9-10: Correlation of age of onset and affected hands**

9.7.5 Discussion

The involvement of both hands at the time of the first consultation is seen in three quarters of the patients. It is not related to a positive family history nor to the age of onset of the disease. However, a statistically significant relation exists with the presence of ectopic lesions.

9.8 Associated diseases

The association of Dupuytren's contracture with other diseases, especially alcoholism, diabetes, and epilepsy, has been recognized for many years (McFarlane 1985a). A summary of the findings in this group of patients is given in figure 13.


	Figure 9-13: Personal medical history and associated disease

9.8.1 Local trauma

A history of local trauma was noted in 9.8% of the patients and was interpreted by many of them as the starting point of the disease.

9.8.1.1 Local trauma and sex of the patients

As shown in table 11, there is no relation between a history of local trauma and the sex of the patients. A positive history was recorded in 10.1 % of the men and 9.3% of the women. That small difference is not statistically significant (p<0.9; Statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-11: Crosstabulation of history of local trauma and sex of the patient.**

9.8.1.2 Local trauma and involvement of both hands

Table 12 shows a clear difference in the distribution of the history of local trauma between patients with one or both hands affected. Among patients with one hand affected, 23.8% had a history of local trauma whereas among patients with both hands affected only 7.9% reported such a history. This difference is statistically very significant (p<0.001; statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-12: Crosstabulation of history of local trauma and involvement of both hands**

9.8.1.3 Local trauma and family history

As shown in table 13, there is no relation between a history of local trauma and a family history of the disease. The frequency of local trauma was 10.1% if the family history was negative and 11.0% if it was positive. That small difference is not statistically significant (p<0.8; Statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-13: Crosstabulation of history of local trauma and family history of the disease**

9.8.1.4 Local trauma and ectopic localizations

As shown in table 14, there is no relation between a history of local trauma and the presence of ectopic localizations of the disease (p<0.6; Statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-14: Crosstabulation of history of local trauma and presence of ectopic sites of the disease**

9.8.1.5 Local trauma and age of onset

A relation was found between the history of a local trauma and the mean age of onset of the disease (table 15). The observed difference is statistically significant (p<0.05; statistical methodology: two-sample separate-variance t test. Levene's test was used to test the hypothesis that the two population variances were equal what they were not).

**Table 9-15: Relation between age of onset of the disease ans a local history of the disease**

9.8.1.6 Discussion

Local trauma was noted in 9.8% of the patients in this series. McFarlane et al. (1990a) found 14% in their multicentric survey. Mikkelsen (1990) reported 23.8% in his epidemiological survey. We cannot forget however that he surveyed a whole population and not only patients seeking surgical advice for their disease.

The fact that there is no relation between a history of local trauma and a family history or the presence of ectopic sites of the disease would suggest that we are dealing with the same population of patients. The relation found with the involvement of both hands and the age of onset could then be interpreted as the fact that the local trauma acts as a trigger mechanism in predisposed patients. This would fit in very well with the diathesis theory of Hueston who states (1990a): "In the absence of a racial predisposition to Dupuytren's disease, any amount of injury to the hand will fail to produce it".

9.8.2 Alcoholism

The first extensive study relating alcoholism and Dupuytren's disease was done by Wolfe in 1956. Since then numerous clinical studies have attempted to assess this relation (Hurst et al., 1990). The question has been approached in two different ways. Some investigators have studied the incidence of alcoholism in populations composed of patients with known Dupuytren's disease. Others have studied the incidence of Dupuytren's disease in populations composed of alcoholics. In these studies, the diagnostic basis for alcoholism has been variable (Hurst et al., 1990). Some have simply recorded the medical history of alcoholism on a yes/no basis (McFarlane, 1985a; Hurst 1986), others have tried to evaluate the daily consumption of alcohol. All those studies (Hurst et al., 1990) have concluded to a higher incidence of Dupuytren's disease in alcoholic patients.

In this study, alcoholism was assessed by questioning and checking with the physiotherapist in charge of the post-operative treatment. As shown in figure 13, 9.8% of the patients were assessed as being alcoholics. This percentage is certainly conservative.

9.8.2.1 Alcoholism and sex of the patients

There is a small difference in the proportions of alcoholics between men and women (table 16). Alcoholism was recorded in 10.4% of the men and 7.5% of the women. That difference is not statistically significant (p<0.4; statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-16: Crosstabulation of alcoholism with sex of the patients**

9.8.2.2 Alcoholism and involvement of both hands

Table 17 shows a clear difference in the distribution of alcoholism between patients with one or both hands affected. Among alcoholics, 95.8 % had both hands affected compared with 73% in non-alcoholic patients. This difference is statistically significant (p<0.02; statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-17: Crosstabulation of alcoholism with involvement of both hands**

9.8.2.3 Alcoholism and family history

Among patients without family history of the disease, 10.1% were assessed as being alcoholics compared with 7.9% among patients with a positive history (table 18). The difference is not statistically significant (p<0.5; statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-18: Crosstabulation of alcoholism and family history**

9.8.2.4 Alcoholism and ectopic localisations

21.5% of the patients assessed as non-alcoholics showed signs of involvement of ectopic sites compared with 33.3% of those coded as alcoholics (table 19). This difference which seems considerable is nevertheless statistically not significant (p<0.3; statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-19: Crosstabulation of alcoholism with the presence of ectopic sites of the disease.**

9.8.2.5 Alcoholism and age of onset

No relation was found between alcoholism and the mean age of onset of the disease (table 20) as the small observed difference is statistically not significant (p<0.7; statistical methodology: two-sample pooled-variance t test. Levene's test was used to test the hypothesis that the two population variances were equal (SPSS¹)).

**Table 9-20: Relation between alcoholism and mean age of onset**

9.8.2.6 Discussion

In their survey, McFarlane et al. (1990a) found 10% of alcoholic patients, about the same percentage as we did (9.8%). There are nevertheless some differences between their findings and ours. They found that the alcoholic group had more family history (we did not) and greater incidence of other areas involved (we did not although there was a greater incidence which was not statistically significant). In agreement with our results, they found that there were more bilateral diseases and that the ages of onset were not statistically different.

An interesting aspect of McFarlane's findings is that there are great differences in the reported prevalence of alcoholism among patients suffering from Dupuytren's disease depending on the countries: 8% in the USA but 15 % in Canada, 6% in Germany and the UK but 12% in France, 5% in Australia... Clearly the diagnosis of alcoholism by surgeons is subjective and this, probably, explains some of the differences between McFarlane's results and those detailed above.

9.8.3 Diabetes

From various studies, Hurst et al. (1990) conclude that the true incidence of Dupuytren's disease in diabetes mellitus probably approaches 40% and that diabetics with Dupuytren's disease rarely need surgery. Yet one large surgical series which based its diagnosis of diabetes on a positive history for clinical diabetes found an unexciting 8% incidence (McFarlane 1985a, McFarlane et al.,1990a). Bergenudd (1993) in his study of the prevalence of Dupuytren's disease among participants in the Malmö Longitudinal Study found no relationship with diabetes (diagnosis based on the 2-hour loading test). In this series, 5.5% of the patients were treated for diabetes mellitus.

9.8.3.1 Diabetes and sex of the patients

There is a small difference in the proportions of diabetics between men and women (table 21). Diabetes was recorded in 6.0% of the men and 3.7% of the women. That difference is not statistically significant (p<0.4; statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-21: Crosstabulation of diabetes ans sex of the patients**

9.8.3.2 Diabetes and involvement of both hands

There were no patients treated for diabetes who had only one hand affected by Dupuytren's disease (table 22). Nevertheless this is not really statistically significant (p<0.1; statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-22: Crosstabulation of diabetes with involvement of both hands**

9.8.3.3 Diabetes and family history, diabetes and ectopic localizations, diabetes and age of onset

No association of any kind was found between these variables.

9.8.3.4 Discussion

The absence of statistical association between diabetes and other factors of importance in Dupuytren's disease is in agreement with the findings of McFarlane et al. (1990a). They also found more bilateral diseases in patients with diabetes mellitus.

9.8.4 Epilepsy

Many studies have shown an association between Dupuytren's disease and epilepsy (Hurst et al., 1990). McFarlane (1985a, 1990a) reports an incidence of 3% in his large multi-centric study. As shown in figure 13, the incidence is this series is 5.0%.

9.8.4.1 Epilepsy and sex of the patients

As shown in table 23, no relation exits between these variables (p<0.7).

**Table 9-23: Crosstabulation of epilepsy and sex of the patients**

9.8.4.2 Epilepsy and involvement of both hands

Only 10% of the epileptic patients had only one hand affected compared with 25.4% of the non-epileptic patients (table 24). This difference is nevertheless not statistically significant (p<0.3; statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-24: Crosstabulation and involvement of both hands**

9.8.4.3 Epilepsy and family history

No statistically significant relation (p<0.3) was found between these variables (table 25).

**Table 9-25: Crosstabulation of epilepsy with family history**

9.8.4.4 Epilepsy and ectopic localisations

As shown in table 26, no relation exits between these variables (p<0.7).

**Table 9-26: Crosstabulation of epilepsy with ectopic lesions**

9.8.4.5 Epilepsy and age of onset

The mean age of onset in the group of epileptic patients was lower than in the non-epileptic group (table 27). This difference is nevertheless not statistically significant (p<0.06; statistical methodology: two-sample pooled-variance t test; Levene's test was used to test the hypothesis that the two population variances were equal (SPSS¹)).

**Table 9-27: Relation between epilepsy and age of onset**

9.8.4.6 Discussion

These findings are in agreement with the conclusions of McFarlane who found also more bilateral disease and an earlier age of onset.

9.9 Previous operation

At the time of the first visit 10.4% of the patients had already been operated (table 28).

The relations between recurrences or extensions of Dupuytren's disease and the different variables studied so far will be analysed later in the following chapters.

**Table 9-28: Previous operation**

9.10 The hands

To evaluate the extent of the disease, the mobility of the metacarpophalangeal and interphalangeal joints as well as the distribution of the involved rays were measured. The assessment formula proposed by Tubiana et al. (1968) was used to give a more synthetic estimate of the contracture. To evaluate the functional impact of the extension deficit, the impairment-of-function assessment of Swanson (1983) was also used. A computer program was developed to easily calculate the impairment percentage from the loss of mobility of the joints. Later, this assessment proved useful to show the consequences of possible postoperative complications.
The number of involved rays and the total loss of extension were calculated from the mobility measurements. In order to estimate the rapidity of the development of the disease and its severity, an index of aggressiveness was computed by dividing the total loss of extension by the reported length of evolution.

To evaluate the possible relations between the extent of the disease and the variables studied so far, only hands operated for the first time could obviously be retained to avoid the risk of introducing the after-effects of previous operations. Only the first hand operated in patients with a bilateral disease was considered to avoid a bias that could be introduced by the over-representation of some variables related to a bilateral disease.

The study reviews the condition of 416 hands in 416 patients.

9.11 Tubiana's grade

The mean Tubiana's grade was 3.4 (Std. Dev. 2.1). The distribution of the number of patients by Tubiana's grade is not normal (fig. 14).


	Figure 9-14: Distribution of Tubiana's grade

9.11.1 Tubiana's grade and sex of the patient

The mean Tubiana's grade in men was 3.6 (SD 2.2) and only 2.8 (SD 1.6) in women. This difference is statistically very significant (p<0.001). Statistical methodology: The difference between two groups was tested by the Mann-Whitney test (SPSS⁴) for all the variables analysed here under.

9.11.2 Tubiana's grade and family history, involvement of both hands or ectopic localisations

The results are summarised in table 29.

**Table 9-29: Tubiana's grade and family history, involvement of both hands or ectopic localisations**

9.11.3 Tubiana's grade and associated diseases

The results are summarised in table 30.

**Table 9-30: Tubiana's grade and associated diseases**

9.11.4 Discussion

Men, alcoholics and patients with an involvement of both hands or of ectopic sites have, statistically, a higher Tubiana's grade. Patients with a positive history of local trauma have a lower grade. Family history, diabetes and epilepsy do not seem to play a role.

9.12 Impairment of function

The mean impairment of function was 7.2 % (Std. Dev. 6.1). As shown in figure 15, the distribution of this variable is not normal. The figure shows an interesting point: 50% of the patients who have been operated had an impairment of function less than or equal to 5% and 80% of the patients had less than 12%. This important point must be kept in mind when choosing the appropriate treatment as we shall see later.


	Figure 9-15: Distribution of impairment of function

The preoperative impairment of function is of course directly related to the total loss of extension and to the Tubiana's grade. The relations with the other variables are thus the same. They are summarised in tables 31, 32 and 33.

**Table 9-31: Impairment of function and sex of the patients**

**Table 9-32: Impairment of function and family history, involvement of both hands and ectopic sites of the disease**

**Table 9-33: Impairment of function and associated diseases**

9.13 Total loss of extension

The mean loss of extension was 100.1 (Std. Dev. 75.7). The remark formulated for the impairment of function applies here too. The relations with the other variables are summarised in tables 34, 35 and 36.

**Table 9-34: Total loss of extension and sex of the patients**

**Table 9-35: Total loss of extension and family history, involvement of both hands and ectopic sites of the disease**

**Table 9-36: Total loss of extension and associated disease**

We have tried to know if there was a relation between the loss of extension and the age of the patient or the length of evolution of the symptoms. No trend is visible on the data plots (fig. 16, 17).


	Figure 9-16: Plot of extension deficit by age of the patients


	Figure 9-17: Plot of total extension deficit by length of evolution of the disease

9.14 Number of involved rays

The mean number of involved rays was 2.1 (Std. Dev. 1.1). In 69.5% of the cases, one or two rays only were involved (fig. 18). In Northern European patients, McFarlane et al. (1990a) found this percentage to be 67%. The involvement was not equally distributed among the five rays (fig. 19) the ring and little fingers being, by far, the most common sites of the disease.


	Figure 9-18: Number of involved rays


	Figure 9-19: Percentage of involvement of each ray. The first web is included with the first ray

The presence of Dupuytren tissue in the first two rays (the radial rays) is usually interpreted as a sign of severe disease (McFarlane et al., 1990a). It was seen in 20% of the cases whereas the ulnar rays were invaded in 98.6% of the patients.

9.14.1 Number of rays and sex of the patients

Men have statistically more rays involved than women (table 37).

**Table 9-37: Number of involved rays and sex of the patients**

9.14.2 Number of rays and family history, involvement of both hands or ectopic localisations

The data are summarised in table 38. Patients with a bilateral disease or ectopic sites of the disease have more rays involved (statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-38: Number of involved rays and family history, involvement of both hands and ectopic sites of the disease**

9.14.3 Number of rays and associated disease

In the same way, as seen in table 39, alcoholic patients have more rays invaded whereas those with a positive history of local trauma have less (statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-39: Number of involved rays ans associated disease**

9.14.4 Number of rays and age of the patient or length of evolution of the disease

As shown in the following plots (fig. 20, 21) no relation is apparent.


	Figure 9-20: Plot of number of involved rays by age of the patients


	Figure 9-21: Plot of number of involved rays by length of evolution of the disease

9.15 Involvement of the radial rays

In order to verify the interpretation of the involvement of the first two rays as a sign of severe disease, the relations between this involvement and the other variables studied so far have been checked.

9.15.1 Involvement of the radial rays and sex of the patients

An involvement of the radial rays was found in 22.3% of the men and 11.5% of the women (table 40). This difference is statistically significant (p<0.02; statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-40: Involvement of radial rays and sex of the patients**

9.15.2 Involvement of radial rays and family history

No relation was found as can be seen from table 41 (p<0.7; statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-41: Involvement of radial rays and family history of the disease**

9.15.3 Involvement of radial rays and ectopic sites

In patients with ectopic sites of the disease, an involvement of the radial rays was found in 40.0% of the cases, in the others in only 13.7% (table 42). This difference is very significant (p<0.001; statistical methodology: Pearson chi-square test of independence (SPSS²)).

**Table 9-42: Involvement of radial rays and family history of the disease**

9.15.4 Involvement of radial rays and bilateral disease

Patients suffering from a bilateral disease had an involvement of the radial rays in 25% of the cases compared with 13.8% in those with only one hand affected (table 43). This difference is nevertheless not significant (p<0.2).

**Table 9-43: Involvement of radial rays and bilateral disease**

9.15.5 Involvement of radial rays and associated disease

No relation was found between the involvement of radial rays and associated diseases. Particularly there is no statistically significant relation with a history of local trauma (p<0.5) or with alcoholism (p<0.3) as show in tables 44 and 45.

**Table 9-44: Involvement of radial rays and local trauma**

**Table 9-45: Involvement of radial rays and alcoholism**

9.15.6 Involvement of radial rays and age of the patient

Patients with an involvement of radial rays were slightly younger (table 46). The small observed difference is not statistically significant (statistical methodology: one-way analysis of variance; Levene's test was used to test the hypothesis that the two population variances were equal (SPSS³)).

**Table 9-46: Correlation between age of the patient and involvement of the radial rays**

9.15.7 Involvement of radial rays and length of evolution

The length of evolution of the disease was significantly shorter in patient with an involvement of the radial rays (table 47). On the contrary, no relation at all was found with the age of onset of the disease (p<0.9; statistical methodology: two-sample separate-variance t test. Levene's test was used to test the hypothesis that the two population variances were equal what they were not).

**Table 9-47: Correlation between length of evolution and involvement of radial rays**

9.15.8 Discussion

The involvement of the radial rays seems indeed to be related to a more severe disease since it is more often encountered in men (who have a more invasive disease as we have seen before) and in patients with signs of ectopic lesions. The statistically longer length of evolution could be compatible with this hypothesis.

9.16 Age of onset, length of evolution and severity of the disease

A statistically significant correlation (r=-0.38, p<0.001; statistical methodology: Spearman rank correlation coefficient which is a nonparametric coefficient based on ranks (SPSS⁷)) exists between the age of onset and the length of evolution of the disease: patients in whom a retraction appeared early consult later in the course of the disease (fig. 22).


	Figure 9-22: Relation between length of evolution and age of onset of the disease

The Tubiana's grade increases with the length of evolution (r=0.3, p<0.001) and the number of involved rays as well (r=0.23, p<0.001).

On the contrary, there is no correlation between the Tubiana's grade or the number of involved rays and the age of onset of the disease.

9.16.1 Discussion

An early onset of the contracture is usually interpreted as an indicator of an aggressive disease (Hueston 1985c, 1990a). We would thus expect a correlation between this variable and the severity of the contracture estimated by the Tubiana's grade and the number of involved rays. This correlation does not exist and this casts some doubts on this interpretation. If an early onset indicated a more active disease, we would also expect a shorter evolution before the first consultation because of a rapidly growing impairment of function. This is not the case.

On the other hand the observed relation between the severity of the contracture and the length of evolution is normal in a slowly progressive disease.

9.17 Summary

We are confronted with a great number of inter-related variables and it becomes quite impossible to get a global view of the situation. The data analysed so far are summarised in two tables, one for the data pertaining to the patients (table 48), the other for the variables related to the hand condition (table 49).

**Table 9-48: Relation between patient variables. A "0" denotes the absence of a statistically significant relation**

**Table 9-49: Relations between hand variables**

9.18 Conclusions

From this lengthy analysis of the variables that characterise this population of surgical patients, several conclusions can be drawn.

In all respects, there is no significant difference between our findings and those of Mikkelsen and McFarlane who have published the two most complete studies so far on the same subject. This observed similarity has several consequences:

this series thus confirms the conclusions of these authors
as far as our recruitment is typical for Belgium, one can say that the Belgian surgeon is faced with the same type of Dupuytren's disease as his colleagues working with patients of Northern European origin (it would be odd to have an atypical recruitment for Belgium and still observe the same population parameters as in other Northern European countries)
being certain that the manifestation of the disease is the same as in other countries, we know we can now safely compare our surgical results with those published by others working on the same population.

If we give a closer look at tables 48 and 49, a number of conclusions can be drawn in relation to the Dupuytren diathesis theory:

men have a greater extension deficit, more involved rays, more involvement of the radial rays and an earlier onset of the disease; even if sex is not related to other patient variables, it clearly plays a role in the aggressiveness of the disease
family history, which unreliability has already been discussed, is related to no other variable except the age of onset and the presence of ectopic sites of the disease; from this data, family history cannot be used as an index of the severity of the disease
the presence of ectopic sites of the disease is positively related to a greater extension deficit, more involved rays, more involvement of the radial rays and of both hands; it is certainly a good index of a stronger diathesis
the involvement of both hands, observed in three quarters of the patients, is related to a greater extension deficit and more involved rays but not to radial rays; it is probably not a good indicator of the severity of the disease as it is seen often and the observed frequency would probably even rise on later follow-up
patients reporting a local trauma have signs of a less aggressive disease as they have a later onset of the disease, less involvement of both hands, a lesser extension deficit and less rays involved; from the available data, it is impossible to know if local trauma, as suggested before, just plays the role of a trigger mechanism bringing the patients earlier to the surgeon or if we are faced with another type of disease; the first option is probably the right one
as stated previously, diabetes is, in the literature, associated with a milder form of Dupuytren's disease and epilepsy with a more aggressive one; my results do not confirm this view
alcoholism, certainly under-estimated in this series, is associated with more involvement of both hands, more extension deficit and more rays involved; it is probably related with a more severe course of the disease
the involvement of the radial rays, seen more often in men and in the presence of ectopic sites, is perhaps an indicator of a more severe disease.

In summary, two factors are clearly related to the severity of Dupuytren's disease, namely sex and the presence of ectopic sites. Three are possibly related, the involvement of radial rays, alcoholism and a history of local trauma. The early onset of the disease, bilaterality and a positive family history that Hueston recognises as signs of a strong diathesis do not seem to be of any use for the measure of severity of the disease.